Manual/Guides

Adding Data

The Add Data menu is the main way to bring layers into Meratyn. It groups sources into Files, Web services, Cloud formats, 3D layers, and Databases. You can also drag files straight onto the map.

Files

ItemNotes
Vector LayerOpens the Add Vector panel (backed by maplibre-gl-vector). Loads GeoJSON, GeoParquet, FlatGeobuf, zipped Shapefile, GeoPackage, KML/KMZ, GML, and other vector formats from a file or URL.
Raster LayerOpens the Add Raster panel (backed by maplibre-gl-raster). Loads GeoTIFF and Cloud-Optimized GeoTIFF (COG) from a file or URL. Multiband COGs support bounded clicked-pixel spectral profiles with PNG/CSV export; single-band rasters keep the existing pixel readout. WASM COG layers auto-zoom to their bounds and preserve 128-bin histogram, ramp, and palette previews across map-style changes.
Delimited Text / Excel LayerLoads CSV/TSV or .xls/.xlsx files from the browser or desktop file picker. Excel workbooks expose a worksheet selector, then use longitude/latitude (or X/Y) columns to build point features.
CAD (DXF/DWG) LayerLoads an AutoCAD DXF or DWG drawing as a vector layer. Choose which drawing layer to load, and enter the source CRS (e.g. an EPSG code) to reproject it — leave blank if the drawing is already in longitude/latitude.
File Geodatabase (GDB)Adds a layer from an Esri File Geodatabase (a .gdb folder). Attribute-only tables are not listed; the selected feature class is read through the local Meratyn server and reprojected to WGS84 (desktop app).
Geotagged PhotosImports geotagged JPEG/HEIC photos as a point layer, placing each one at its EXIF GPS location.
GPX LayerLoads a GPX file or URL and splits it into separate waypoint, track, and route layers.
Encoded PolylinePaste Google/OSRM (precision 5) or Valhalla (precision 6) encoded polylines, optionally separated by a custom delimiter, and add them as a line layer.
MBTiles LayerLoads a local MBTiles tile archive (desktop app).

Vector files are reprojected to EPSG:4326 on load. In the browser, vector import relies on DuckDB-WASM Spatial, with direct handling for GeoJSON, zipped Shapefiles, and KMZ archives. To keep the UI responsive, GeoJSON/JSON/KML files at or above 100 MiB and Shapefiles whose uncompressed .shp member reaches 100 MiB bypass the synchronous JavaScript readers and use DuckDB instead. The Add Vector panel uses these same thresholds when it chooses its tiled renderer, so loading a file from the panel and dropping it on the map has the same memory profile. DuckDB still materializes the resulting FeatureCollection, so this is a responsiveness guard rather than unbounded streaming; GPX and CSV/TSV keep their specialized readers. Large standalone KML fallbacks omit ground overlays and 3D models because those require decoding the complete XML document. KML placemarks inside <Folder> elements are imported as named layers in matching nested layer groups; placemarks outside folders remain in one flat layer. Folder splitting is available on the styled XML reader path, while very large KML files routed directly to DuckDB retain a single flat layer.

Raster and cloud-format notes

Meratyn’s COG reader handles tiled JPEG GeoTIFFs (including shared JPEG tables and YCbCr photometric data) through a bounded window cache, so only the requested map windows are decoded. Browser GeoTIFF-to-COG conversion warns at 40 million decoded samples and refuses more than 100 million; use gdal_translate or rio cogeo for larger files. GeoParquet imports read the file’s geo metadata to select the primary geometry column and source CRS (authority code, WKT, or PROJJSON) before reprojection. Files without a declared geometry column can use conventional longitude/latitude or x/y columns to synthesize WGS84 points, and native GeoArrow logical CRS metadata is honored when file-level geo metadata is absent. A single-band classified or custom color-ramp result uses the GPU raster engine automatically; RGB layers and ordinary continuous ramps keep the selected engine.

Web services

ItemNotes
XYZ LayerA raster or vector tile service using a {z}/{x}/{y} URL template.
WMS LayerA Web Map Service layer, with click-to-identify through GetFeatureInfo where supported.
WFS LayerA Web Feature Service layer, with optional automatic refresh.
WMTS LayerA Web Map Tile Service layer.
ArcGIS LayerAn ArcGIS FeatureServer, VectorTileServer, MapServer, or ImageServer layer (or a portal item).
STAC URL / COG LayerOpens a specific STAC catalog or API URL and adds the selected COG asset as a raster layer. For catalog discovery and broader asset support, use the separate STAC Catalogs plugin.
STAC CatalogsActivates the catalog browser, which discovers public catalogs from STAC Index, searches collections and items, and adds supported assets. This is separate from the URL-based STAC / COG loader above.
OGC API - FeaturesAdds a collection from a service landing page, /collections, or an items URL. Choose a page size and optional bounding box or date filter. Paging continues until the service is exhausted or the 64 MiB loading budget is reached; incomplete results are identified.
OGC Vector TilesAdds an OGC API - Tiles (vector) service from a TileJSON URL or {z}/{y}/{x} MVT tile template. An optional Mapbox/MapLibre style URL (or comma-separated source layer names) supplies the tileset’s source layers.
GeoRSS LayerAdds a GeoRSS feed (RSS, Atom, or RDF) from a URL or file as a vector layer.
Video LayerA georeferenced video overlay: supply an MP4 URL plus four corner coordinates to place the video on the map.
Deck.gl LayerPick a deck.gl layer type, then load its example data or your own CSV/JSON/GeoJSON from a file or URL. Large datasets are stored inline in the project file, so oversized uploads warn before loading.

WFS paging

WFS layers load the current map extent. Features per page defaults to 1,000; it is the request size, not a total feature limit. Moving the map loads the new extent, including views across the antimeridian. Requests use reusable cells inside the layer’s advertised bounds, with smaller cells for dense areas. Completed cells are cached for five minutes within a shared memory budget. Returning to a cached area reuses its data; expired or evicted cells are fetched again. Without advertised bounds, exact requested extents are cached instead. Refresh invalidates cached coverage and reloads the current view. WFS and OGC feature downloads share two request slots, releasing a slot between pages. The attribute table identifies its extent coverage. A loading budget, failed page, or service that ignores pagination leaves an explicit partial-data warning. When a layer holds more than the memory budget, the features in view are kept and features outside it are released, so panning or zooming in loads the area you are looking at; features already on the map are not downloaded again into the budget. Analysis that requires the complete dataset does not silently use this subset.

ArcGIS services

The ArcGIS source accepts service URLs or portal item IDs. FeatureServer layers are downloaded as GeoJSON; VectorTileServer layers use their published style; MapServer and ImageServer layers become ordinary raster layers. A standard Web Mercator MapServer cache is consumed as {z}/{y}/{x} tiles, while other map services use /export and ImageServer uses /exportImage per tile. MapServer sublayers can be selected by numeric id (including by browsing the service), and ImageServer raster functions can be selected from the service catalog or entered as a custom JSON rendering rule. Tokens remain in memory and are removed from projects shared or exported from Meratyn.

FeatureServer downloads are requested in bounded pages. When the map is available, an added FeatureServer layer instead follows the current viewport: each pan or zoom replaces the previous request, splits antimeridian-crossing views into two envelopes, and publishes pages as they arrive. Stale or cancelled requests cannot overwrite a newer view. Project restore rebinds this viewport loader, while manual refresh replays the saved paging policy for a headless/full-layer load.

Cloud formats

ItemNotes
GeoParquet LayerCloud-native columnar vector format. Opens the same Add Vector panel as Vector Layer, selects the primary geometry and declared CRS from geo metadata (or native logical metadata), and can synthesize points from conventional longitude/latitude or x/y columns. Can be streamed in place with HTTP range requests for large remote files.
FlatGeobuf LayerCloud-optimized vector format with spatial indexing.
PMTiles LayerA single-file vector or raster tile archive.
Zarr LayerChunked, cloud-native multidimensional arrays.
NetCDF / HDFOpens local or range-readable remote NetCDF/HDF data and adds a raster layer. Scientific cubes expose bounded band and dimension exploration, pixel sampling, and PNG/CSV spectral-profile export.

3D layers

Hyperspectral and 3D cubes

NetCDF/HDF cube exploration runs in the browser only within a bounded envelope: grids up to 192 × 192, at most 64 bands, and no more than 12 MiB of decoded sample data per inspection. The panel reads one requested window at a time and releases it with the panel lifecycle, so it does not create a second map renderer or a continuously repainting GPU loop. If a dataset exceeds any bound, Meratyn keeps the regular 2D NetCDF/HDF path available and explains that the 3D cube view was declined for safety. Remote files must support byte ranges; kerchunk manifests remain supported for cloud-optimized inputs.

ItemNotes
LiDAR LayerPoint-cloud visualization, rendered with deck.gl.
Splatting LayerGaussian splat scenes.
3D Tiles LayerOGC 3D Tiles, restored when reopening a project. Includes a Google Photorealistic 3D Tiles sample that works when a Google Maps API key has been configured for your deployment.
3D Model (glTF)Places a glTF/GLB model at one coordinate or at point features through deck.gl.

Databases

ItemNotes
DuckDB LayerQuery a DuckDB or DuckDB Spatial source and add the result as a layer, with identify, selection, and attribute table support.
PostgreSQL LayerAdd a layer from a PostgreSQL/PostGIS connection (desktop app, served through a local tile server). Editable tables with multiple registered geometry columns expose a geometry-column selector; edits and write-back use the selected column.
Iceberg LayerInspect a credential-free Apache Iceberg REST catalog or table and browse its rows in pages from the Layers panel. The default page size is 5,000 rows, with a 64 MiB decoded-data budget per page. Refresh reloads the current page. Credential-bearing locations and signed URLs are rejected. Direct metadata above a reported 100 MiB is rejected; this preflight cannot measure aggregate manifest/data reads.

Drag and drop

Drag a vector file (GeoJSON, zipped Shapefile, KMZ, and similar) or a GeoTIFF/COG raster directly onto the map to add it as a layer. GPX files dropped on the map are split into named waypoint, track, and route layers.

Basemaps

The basemap sits at the bottom of the Layers panel as the Background entry. Activate the Basemaps plugin from the Plugins menu to switch between OpenFreeMap styles (Liberty, Positron, Bright, Dark, Fiord, 3D), a blank background, or a custom style URL. You can toggle basemap visibility and adjust its opacity from the Layers panel.

Other celestial bodies

Meratyn can map worlds beyond Earth. The Change background dialog and the New project dialog group planetary basemaps into sections for The Moon, Mars, and a collapsible Other celestial bodies section covering Mercury, Venus, the Galilean moons (Io, Europa, Ganymede, Callisto), Titan, Pluto, and Charon. The Moon and Mars mosaics come from OpenPlanetaryMap; the other bodies come from USGS Astrogeology and are reprojected to Web Mercator on the fly so MapLibre can render them.

For quick switching, use the planet switcher (the orbit icon) in the Layers panel header. Selecting a body sets the project’s ellipsoid, so distance, area, and scale-bar measurements use that body’s radius instead of Earth’s. Turf-based buffers, cell-coverage circles, movement analysis, GPS tracks, print scale bars, and elevation profiles apply the same radius correction; the terrain Measure panel identifies non-Earth results as a spherical approximation.

More data sources

Additional catalogs and providers are available as panels and plugins rather than Add Data items, including Planetary Computer, Earth Engine, Overture Maps, and several federal Web Services. See Data Integrations.

Browser vs desktop URL-based sources and the listed browser file imports work in both the browser and desktop app. Local MBTiles, File Geodatabases, PostgreSQL, and native-filesystem-only workflows require the desktop app.